An approach to improve direct runoff estimates and reduce uncertainty in the calculated groundwater component in water balances of large lakes

Abstract

Summary Groundwater is important in the overall water budget of a lake because it affects the quantity and quality of surface water and the ecological health of the lake. The water balance equation is frequently used to estimate the net groundwater flow for small lakes but is seldom used to determine net groundwater flow components for large lakes because: (1) errors accumulate in the calculated groundwater term, and (2) there is an inability to accurately quantify the direct runoff component. In this water balance study of Lake Pyhajarvi (155 km 2 ) in Finland, it was hypothesized a hydrograph separation model could be used to estimate direct runoff to the lake and, when combined with a rigorous uncertainty analyses, would provide reliable net groundwater flow estimates. The PART hydrograph separation model was used to estimate annual per unit area direct runoff values for the watershed of the inflowing Ylaneenjoki River (a subwatershed of the lake) which were then applied to other physically similar subwatersheds of the lake to estimate total direct runoff to the lake. The hydrograph separation method provided superior results and had lower uncertainty than the common approach of using a runoff coefficient based method. The average net groundwater flow into the lake was calculated to be +43 mm per year (+3.0% of average total inflow) for the 38 water years 1971–2008. It varied from −197 mm to 284 mm over that time, and had a magnitude greater than the uncertainty for 17 of the 38 years. The average indirect groundwater contribution to the lake (i.e., the groundwater part of the inflowing rivers) was 454 mm per year (+32% of average total inflow) and demonstrates the overall importance of groundwater. The techniques in this study are applicable to other large lakes and may allow small net groundwater flows to be reliably quantified in settings that might otherwise be unquantifiable or completely lost in large uncertainties.